Modeling and analysis of a cable-driven serial-parallel manipulator

Abstract

Cable-driven serial-parallel (CDSP) manipulators have a small size and large load/mass ratio and represent an ideal scheme for humanoid manipulators. However, owing to the multilevel mapping relationship among the end-effector, joints, cables, and motors, as well as the complex cable transmission path and cable–pulley contact, the modeling and control of such manipulators are highly complex. In this study, the kinematics and dynamics models for a CDSP manipulator were established, and a dynamic controller was proposed. First, the structure and cable transmission path of a CDSP manipulator were described. A kinematics equivalent method was used to equate the cable-driven parallel joint in the manipulator to an Equivalent pair. Using the proposed equivalent method, along with analytical and numerical methods, the kinematic equations were derived. In addition, the dynamic equations including those for the cable–pulley contact, cable tension, and joint variables were derived. Considering the actual control requirements, the cable tensions were optimized to obtain the minimum solution. Moreover, a dynamic control strategy was formulated. Finally, a prototype of the proposed manipulator was fabricated, and the cable tension of its joints and the pose accuracy of its end-effector were tested to verify the models and dynamic control method. The experiment results demonstrated that the kinematic and dynamic models are valid, and the dynamic controller is effective.

Utilizing Reinforcement Learning to Drive Redundant Constrained Cable-Driven Robots with Unknown Parameters

“…The cable-driven robot has garnered considerable attention from researchers owing to its impressive performance [1,2], which includes a high load-to-weight ratio, potential for a large forming space, high-speed motion, and ease of reconstruction [3][4][5][6][7][8]. Additionally, the cable-driven method offers significant flexibility compared to other drive mechanisms.…”